# Regarding run\_t5\_mlm\_flax.py

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fregarding-run-t5-mlm-flax-py%2F42512%3E%5C
**Category:** Beginners\
**Created:** [June 8, 2023, 8:03am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fregarding-run-t5-mlm-flax-py%2F42512 "2023-06-08T08:03:19Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![vrmarathe](https://avatars.discourse-cdn.com/v4/letter/v/cc9497/32.png) [@vrmarathe](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fvrmarathe)%5C
**Post date:** [June 8, 2023, 8:03am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fregarding-run-t5-mlm-flax-py%2F42512%2F1 "2023-06-08T08:03:19Z")

</div>

I am trying to run the example code run\_t5\_mlm\_flax.py. I have the following GPU configuration when I run nvidia-smi

 ![image](https://us1.discourse-cdn.com/hellohellohello/original/3X/e/6/e6208f873e552b378a14d991a0cf44bf4bc05444.png)

Questions:  
1.Could I know which JAX,Flax versions do I need to install from the JAX github repo ?  
2. I am also getting the following errors:

2023-06-08 01:38:47.110150: E external/xla/xla/stream\_executor/cuda/cuda\_dnn.cc:417] Loaded runtime CuDNN library: 8.4.1 but source was compiled with: 8.6.0. CuDNN library needs to have matching major version and equal or higher minor version. If using a binary install, upgrade your CuDNN library. If building from sources, make sure the library loaded at runtime is compatible with the version specified during compile configuration.

**I downgraded the JAX,Flax to a lower versions to cudnn 802 then I got the following error**

2023-06-08 01:23:26.145096: W external/org\_tensorflow/tensorflow/stream\_executor/gpu/asm\_compiler.cc:111] \*\*\* WARNING \*\*\* You are using ptxas 11.0.194, which is older than 11.1. ptxas before 11.1 is known to miscompile XLA code, leading to incorrect results or invalid-address errors.

You may not need to update to CUDA 11.1; cherry-picking the ptxas binary is often sufficient.  
2023-06-08 01:23:26.149641: W external/org\_tensorflow/tensorflow/stream\_executor/gpu/asm\_compiler.cc:230] Falling back to the CUDA driver for PTX compilation; ptxas does not support CC 8.6  
2023-06-08 01:23:26.149664: W external/org\_tensorflow/tensorflow/stream\_executor/gpu/asm\_compiler.cc:233] Used ptxas at ptxas  
2023-06-08 01:23:26.152107: E external/org\_tensorflow/tensorflow/stream\_executor/cuda/cuda\_driver.cc:632] failed to get PTX kernel “shift\_right\_logical\_3” from module: CUDA\_ERROR\_NOT\_FOUND: named symbol not found

Any solutions would be greatly appreciated. I am doing currently a test run on the sample example provided, then I will use it pre-train on my own dataset which I have processed.